Contract documents and guidelines
Here you find contract documents for using Oral-History.Digital. A selection of recommendations, sample templates and interesting links for creating, transcribing, archiving and curating interview collections can be found on the German "Dokumente" page.
Contract documents
- Contract for a scientific service (pdf, sample)
- Annex 1: Terms and conditions (pdf)
- Annex 2: Cost table (pdf)
- Annex 3: Archive description (docx, fillable)
- Agreement on joint responsibility for data processing ("Joint Control", pdf, sample) with annex "TOM" for archive holders (docx, fillable)
- Contact details form for the conclusion of the contract
The software pf oral.history-digital is available as open source via https://github.com/oral-history-digital/ohd.
Media Management Tool (MMT)
MMT supports researchers at Freie Universität Berlin (FUB) and other research institutions in processing audiovisual, multimodal research data. It facilitates the transfer of data – including large media files – their transcoding into user-friendly formats, automatic transcription, multimodal anonymisation and content indexing for academic purposes. The MMT is operated by the University Library of Freie Universität Berlin as part of Oral-History.Digital. All data is processed exclusively on FUB servers, even when AI-supported tools are used. Use of the MMT requires users to be authorised by the FUB. To do so, users must agree to the MMT’s terms and conditions.
Access to MMT is provided on request by emailing mmt@oral-history.digital.
Automatic Speech Recognition (ASR)
Automatic Speech Recognition (ASR) can produce a usable raw transcript – complete with detailed timecodes – provided the recording quality is good and the pronunciation is clear in common languages. In many cases, however, manual (post-)transcription is required. The ASR4Memory project has developed a service for the automatic transcription of audiovisual research data from the historical humanities. ASR4Memory uses WhisperX technology, which currently supports transcription in 30 languages (list of language models). Before speech recognition takes place, the audio frequency ranges are automatically optimised. The generated transcripts are tagged with word- or sentence-based timestamps and speaker identifiers. Various export formats (including csv, vtt and TEI-xml) allow for direct import into the Oral-History.Digital cataloguing platform, as well as manual post-processing in a transcription programme. In accordance with data protection regulations, the data is processed exclusively on locally operated infrastructure at Freie Universität Berlin. The service is accessible via the Media Management Tool (MMT) of Oral-History.Digital.
Recommendations
The German documents at the "Dokumente" page cover topics like funding, data management plans, interviewing, recording techniques and media formats, consent forms, risk assessment, transcription, speech recognition, alignment, indexing, topic modeling, software and more.
Some useful information in English can be found at
